Steve,
When I wrote:
> <<I don't agree that we have no clues at all. Firstly, the 09 lineage had
> quite recent (just one node back) ancestors with three groups that are
> dominant (or it's close to be dominant) in Near East and northern Africa
> (the "89 only" group), in Caucasus and Near East (the 172 lineage), in
> Europe (the 170 lineage) and one group that is rare and present mostly in
> Southern Asia (the 52 + 69 group).>>
I forgot to put the word "common" before the word "ancestors". It could be
deduced from my previous messages anyway (that I thought about "common
ancestor" of 09, 89, 172, 170 and 52/69 group in this case), but,
nevertheless, it was my fault, so please accept my apology for confusing
you. I hope you'll forgive me that I will not comment the part of your
message that is related to this misunderstanding.
>The ONLY
> thing that a phylogenic tree tells you is the necessary order of events,
but
> nothing about absolute time.
The order of events is the only thing that can be said with almost complete
certainity. In addition to that, it is also possible to roughly asses the
relative distance between different nodes (though the reliability of such
analysis may vary substantially, depending for example on how the analysis
was performed). You are right, however, that it says almost nothing about
absolute time (but it can be correlated with other type of data; see the
example of migration to America).
> And in the case of two parallel branching
> decendents, it will not even tell you which one branched off first, or the
> time interval between the branchings. And the odds are always they were
not
> simultaeneous.
I completely agree. The "true" simultaneous branchings are not very likely.
We can never be certain that a nod leading to 4-6 separate branches
corresponds to simultaneous multiple splits or maybe these are sequential
splits. (Much more likely is that there was some sequential braching). On
the other hand, since no new mutations (that would be common to subsequently
spliting branches) appeared between these sequential splits, we can suspect
that either the original population was relatively large (so the new
mutations could not effectively dominate the population) or the subsequent
splits were relatively quick (if not simultaneous).
> In fact, there is no way in telling in a phylogenic tree whether M172
arose
> 5000 years before M09 or 5000 years after. (Ave mutations from the root
> might justify some inference re absolute time, but this is complicated by
the
> fact that mutation rates ordinarily vary by 1000's of years.) On a tiny
> scale, it's like the situation I heard of where a man had a grandchild who
> was older than his brother. In deep time lineage, this sort of thing
happens
> more often than not. Some of these existing anchor haplotypes are in
theory
> 10's of 1000's of years old and could have mutated different descendents
at
> vastly different times.
Again, this is all very true. I expect, however, that in most such
"asymetric" cases, the symetricity will be at least partially regained, as
some of the populations (and haplotypes) will be studied more thoroughly.
For example, when you have a hyplotype that is very rare (and limited to one
terytory) you do not try to find subsequent polymorhisms defining subsequent
branches, since what you have is distinctive enough. But when you find a
mutation that defines a new very common haplogroup, you will try to find
additional (subsequent or preceding) polymorhisms that will better
characterize this large population of chromosomes. However, the frequency of
new mutations appearing in the lineage may sometimes mean something very
important (at least this is my opinion). Knowing how the genetic drift
works, we could expect that new mutations have better chance to exclude the
nonmutated allele when the population is small. When you have two sister
branches, one of which includes seven subsequent mutations whereas the other
is characterized by lack of any mutation that would additionaly distinguish
it from other sister branches (compare the hyplotype containing mutations
89, 09, 45,74, 173, 17, 64 and 87,and the "89 alone" hyplotype) it seems
possible that the former could evolve while beeing carried by very small
populations, whereas the latter was established and maintained in pretty
large population. This is, however, only my guess (and I'm not particularly
strong in population genetics). What is your opinion in this matter?
> All we really know is that each of these two mutations (09 and 172)
happened
> in different individuals somewhere, and both of these individuals were
> carrying M89 traits (haplotype 71), but neither was carrying ANY parallel
or
> later mutations (e.g., 170, 62, 52 or 69.) If you look at Table 2 in
> Underhill, you'll see for example that the last "defining mutation" of
> Haplogroup VI (yellow) is nothing more than the presence of M89 but the
> absence of M09.
Right. And I think they should avoid creating such "artificial" haplogroups
(this is even more confusing in case of haplogroup VIII).
> So, the haplotypes or mutations you mentioned above don't tell us where or
> when 09 happened or if it did or did not predate those mutations you
mention.
These are things that for the reasons you described (and not only those)
cannot be certain. It doesn't , however, mean that we cannot establish the
most likely scenario, that would lead to what we observe today. As for the
09 mutation, the location of sister branches that evolved from the common
ancestor (characterized by the new 89 mutation) indicates that it couldn't
be very far from Near East (do you have the more likely alternative?).
Additionally, the location of haplotypes evolving from the 09 progenitor
suggests that Southern or Central Asia was a most likely place were most of
the descendant branches split off. (What is your opinion about the more
likely alternative in this case?)
Regards,
Michal